Data

Tools

Indexes

Applications

Experiments

Open Source Science

https://github.com/bowang-lab/concerto

Carcinogenicity prediction with graph neural networks

https://github.com/bowang-lab/concerto

Science Score: 26.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • ○
    CITATION.cff file
  • ✓
    codemeta.json file
    Found codemeta.json file
  • ✓
    .zenodo.json file
    Found .zenodo.json file
  • ○
    DOI references
  • ○
    Academic publication links
  • ○
    Committers with academic emails
  • ○
    Institutional organization owner
  • ○
    JOSS paper metadata
  • ○
    Scientific vocabulary similarity
    Low similarity (7.0%) to scientific vocabulary
Last synced: 11 months ago · JSON representation

Repository

Carcinogenicity prediction with graph neural networks

Basic Info
  • Host: GitHub
  • Owner: bowang-lab
  • License: mit
  • Language: Jupyter Notebook
  • Default Branch: main
  • Size: 27.5 MB
Statistics
  • Stars: 10
  • Watchers: 4
  • Forks: 3
  • Open Issues: 0
  • Releases: 0
Created over 4 years ago · Last pushed over 3 years ago
Metadata Files
Readme License

README.md

A Graph Neural Network Approach to MoleculeCarcinogenicity Prediction

alt text

Abstract:

Molecular carcinogenicity is a preventable cause of cancer, however, most experimentaltesting of molecular compounds is an expensive and time consuming process, making highthroughput experimental approaches infeasible. In recent years, there has been substantialprogress in adopting machine learning as an appealing alternative forin vivoprediction ofmolecule carcinogenicity. In this work, we propose a new system, CONCERTO, which usesa graph transformer in conjunction with a molecular fingerprint representation, trained onmulti-round mutagenicity and carcinogenicity objectives. To train and validate CONCERTO,we augment the training dataset with more informative labels and utilize a larger externalvalidation dataset. Extensive experiments demonstrate that our model yields results superiorto alternate approaches for molecular carcinogenicity prediction.

Reproduce:

Code

To train CONCERTO:

First you need to download the GROVER repo from https://github.com/adamoyoung/grover and download the model files as indicated under Pretained Model Download header. Then you should be able to train the model with the hyperparameters described below.

train.py --atom_pairs_fingerprints=false --batch_size=512 --carc_percentile_to_drop=1 --cross_validation=true --ff_dropout=0.1 --ff_hidden_feats=64 --ff_num_layers=5 --fp_nbits=2048 --gradient_clip_norm=10000 --grover_fp=large --lr=0.00044800187354224327 --lr_decay_factor=0.5 --min_carbon_count=3 --model_type=mlp_fingerprint --mut_pre_training=true --network_weight_decay=0.01 --num_mut_pre_training_loop=2 --patience=20 --torsion_fingerprints=false

To conduct counterfactual analysis:

genetic_explain.py 1svdygk7 --wandb_mode online --lower_bound_percent 0 --upper_bound_percent 100 --delta 0.5 --num_generations 1 --num_mols_to_explain 1000 --num_viz 150 --num_samples 3000 --explain_name lots_of_mols

Model Files

Model files and data results are located in: ./CONCERTO_model_and_results/

To load model

``` configfn = "config0.json" with open(os.path.join(newrundp, configfn),'r') as configfile: args = json.load(configfile) args["device"] = dev trainloader, valloader, testloader, heldouttestdataloader, datafeats = loaddata(args)

model = getmodel(args,datafeats)

model.to(dev) checkpoint = th.load(os.path.join(newrundp,checkpointfn),maplocation=dev) model.loadstatedict(checkpoint)

```

Owner

  • Name: WangLab @ U of T
  • Login: bowang-lab
  • Kind: organization
  • Location: 190 Elizabeth St, Toronto, ON M5G 2C4 Canada

BoWang's Lab at University of Toronto

GitHub Events

Total
Last Year

Committers

Last synced: over 1 year ago

All Time
  • Total Commits: 6
  • Total Committers: 2
  • Avg Commits per committer: 3.0
  • Development Distribution Score (DDS): 0.167
Past Year
  • Commits: 0
  • Committers: 0
  • Avg Commits per committer: 0.0
  • Development Distribution Score (DDS): 0.0
Top Committers
Name Email Commits
Phil F p****n@g****m 5
Philip Fradkin p****n@P****l 1

Issues and Pull Requests

Last synced: over 1 year ago

All Time
  • Total issues: 1
  • Total pull requests: 0
  • Average time to close issues: 1 day
  • Average time to close pull requests: N/A
  • Total issue authors: 1
  • Total pull request authors: 0
  • Average comments per issue: 2.0
  • Average comments per pull request: 0
  • Merged pull requests: 0
  • Bot issues: 0
  • Bot pull requests: 0
Past Year
  • Issues: 1
  • Pull requests: 0
  • Average time to close issues: 1 day
  • Average time to close pull requests: N/A
  • Issue authors: 1
  • Pull request authors: 0
  • Average comments per issue: 2.0
  • Average comments per pull request: 0
  • Merged pull requests: 0
  • Bot issues: 0
  • Bot pull requests: 0
View more stats
Top Authors
Issue Authors
  • iehaoang (1)
Pull Request Authors
Top Labels
Issue Labels
Pull Request Labels